A Highly Stable and Sustainable Low-Temperature Selective Absorber: Structural and Ageing Characterisation

Solar absorbers in a three-layer configuration have been prepared by dip-coating onto aluminium substrates. They are constituted by two spinel layers with one silica layer on the top and values of solar absorptance above 0.950 and thermal emittance below 0.04 were obtained. The effects of using different sintering conditions of the upper silica layer on the optical behaviour and durability tests have been studied. Results obtained in accelerated ageing methods, such as thermal stability tests and condensation tests, clearly show that the proposed selective absorber exhibits excellent thermal stability and very good humidity resistance. The results show that the protective action is due not only to the silica layer but also to the alumina layer produced during the absorber preparation. The phase composition of the individual layers was independently confirmed using X-ray diffraction and corroborated by X-ray Photoelectron Spectroscopy. Spinel-like phases were obtained in both the first and second layers. The ageing study shows that the three-layer configuration proposed has a very high potential, in terms of both durability and optical behaviour, for solar thermal low-temperature applications.


Introduction
Energy demands and global warming are increasing as a consequence of the use of conventional energy sources (fossil fuel, gasoline, gas and charcoal) that satisfy around 80% of the global demand for energy [1]. During the last few years, the total carbon dioxide emissions have increased by 54% demonstrating the dominance of fossil fuels worldwide [2,3]. Given the current energy crisis and the critical environmental concerns as exposed in the Paris Agreement in 2015, it is indispensable to promote the development of sustainable sources of energy for reducing polluting emissions, increasing renewable electricity production and improving energy efficiency. This decarbonisation of the economy would be achieved with the use of renewable energies. Solar thermal energy, which uses the full-solar spectrum without greenhouse gas emissions for producing heat, can be a potential candidate to face this universal issue by replacing fossil fuels.
Focusing on low and medium temperature solar applications, several new designs have been developed during the last three decades in order to enhance the thermal performance of flat plate collectors, which are widely used in water and space heating [4]. The success of this technology relies on the solar absorber material that has to combine high optical efficiency with long-term performance stability. The optical requirement that the material must accomplish is a high selectivity, which is achieved when a high solar absorptance (α s ) in the solar wavelength range (0.3-2.5 µm), and a low thermal emittance (ε T ) in the mid/far-infrared wavelength ranges (>2.5 µm) is realised [5]. Solar selective coatings, specifically those based on spinel-type mixed oxides, are widely used in the field of solar thermal energy conversion in order to ensure a greater output of the collector. Apart from the well-known field of selective absorbers, spinel-type mixed oxides have been reported to be suitable for several catalytic processes, thermochemical energy storage and protective coating applications, among others [6][7][8]. This growing interest in spinel-type transition metal oxides in many research fields is due to their broad variety of physical and chemical properties.
Spinel structures have the possibility of substituting a large number of transition metals and of replacing the cations present in the spinel crystal lattice, in both tetrahedral and octahedral sites, which allows chemical modifications and therefore the tuning of their properties [9,10]. This can be achieved as the physicochemical properties of the spinel-structure compounds rigorously depend on the type, charge, and distribution in tetrahedral/octahedral sites of the cations [11,12]. This remarkable peculiarity in terms of the huge versatility of the spinel coatings as well as their inherent high resistance to oxidation and high temperature stability has generated special interest in the field of spectrally selective absorber coatings during the last decades [13][14][15].
This current interest is reflected in our previous published work in which a multilayered absorber for low-temperature thermal applications based on CuMnO x and two antireflective coatings deposited by the dip-coating method on an etched aluminium substrate (Al-CuMnO x /CuFeMnO x /SiO 2 ) was reported [16]. The good spectral selectivity achieved, with solar absorptance of 0.957 and thermal emittance at 100 • C of 0.038, demonstrated the possibility of this three-layer configuration selective absorber to be considered as a future competitive material by the industry. However, before further development or commercialisation, it is critical to evaluate the durability of the solar absorber in terms of performance during the collector's lifetime. The service life of a material depends on several variables which influence its behaviour under service conditions such as its physical and chemical properties, its performance requirement in the application considered, and on external environmental factors. Consequently, apart from high spectral selectivity (high absorptance and low emittance in the range of 0.3-2.5 µm and 2.5-17 µm, respectively), solar selective coatings have to possess high chemical and thermal stabilities at typical field operation temperatures or under regular variations in operating temperature [17].
In the case of low-temperature solar collectors, as they operate in outdoor conditions, it is indispensable for these systems to withstand external environmental factors such as high humidity, high temperatures, atmospheric corrosion and ultraviolet irradiance or wind and snow loads depending on the geographic position [18][19][20]. Focusing on the selective absorber coatings composed of inorganic oxides, high temperatures can speed up oxidation processes and high levels of humidity may create hydrolytic reactions (electrochemical corrosion) further oxidating the surface [21][22][23][24][25]. It should be taken into account that each of these manifold micro-climatic influences can degrade the absorber and the rest of the components of a collector, causing the absorber's optical properties to deteriorate and the loss of the absorber performance over the years [26,27].
In this work, the influence of the SiO 2 layer sintering on the optical parameters and long-term durability of the proposed Al/CuMnO x /CuFeMnO x /SiO 2 selective absorber is reported. The structural analysis of each constituent layer of the selective absorber by X-ray diffraction (XRD) and X-ray photoelectron spectroscopy (XPS) is also included.

Sample Preparation
A chemical etching treatment (10 s in a 5% HF solution) of the aluminium substrate was carried out before depositing the constituent absorber films. Next, each layer of the selective absorber (CuMnO x /CuFeMnO x /SiO 2 ) was separately deposited by dip-coating at 22 • C. All the solutions were kept at this temperature during the absorber preparation to ensure a constant viscosity. The CuMnO x and CuFeMnO x solutions were obtained by dissolving the metallic precursors Cu(NO 3 ) 2 (Sigma Aldrich, St. Louis, MO, USA, 98-103%), Mn(NO 3 ) 2 (Panreac, 97%) and Fe(NO 3 ) 3 (PRS Panreac, 98%) in absolute ethanol at molar ratios equal to 1:1 for CuMnO x and 2:1:1 for CuFeMnO x . In order to stabilise these precursor alcoholic solutions and improve the adhesion of liquid-films, some reagents such as acid, complexing agent and wetting additive were added [28]. Regarding the SiO 2 anti-reflective coating, the solution was prepared via sol-gel by mixing the precursor Tetraethylortosilicate (TEOS) with distilled water and ethanol in a molar ratio of 1:18:5 [16] with hydrochloric acid added as catalyst.
The three layers deposition was performed at the optimised withdrawal rates of 42 cm/min for the CuMnO x film, 18 cm/min for the first CuFeMnO x antireflective coating and 12 cm/min for the SiO 2 anti-reflective coating. These rates were selected as optimal after carefully studying the effect of the withdrawal rate and, therefore, the thickness of each constituent layer in the optical properties (high solar absorptance and low thermal emittance) [16]. The resulting layers of CuMnO x and CuFeMnO x were annealed separately at 600 • C. Regarding the SiO 2 layer, different sintering conditions (blower or oven, 5 or 15-30 min) were tested at 500 • C in order to study the possible influence of this parameter on the complete solar absorber after long term durability tests. Moreover, the condensation resistance of the SiO 2 layer sintered at the different conditions was also studied independently by preparing this layer directly on the aluminium substrate.

Sample Characterisation
A Pananlytical XPERT X-ray diffraction spectrometer (XRD) was used for characterising the constituent absorber thin films. The structural analysis was carried out by identifying the crystalline phases composing the coatings. Additionally, a Perkin-Elmer PHI 5400 X-ray photoelectron spectrometer (XPS) equipped with a Mg Kα excitation source (hν = 1253.6 eV) and a beam size of 1 mm diameter was used to determine the composition of the films and to identify the valence state and cation distribution of the components present in each layer. The intensities were estimated by calculating the area under each peak after smoothing and removing the background using the modified method of Shirley and fitting the experimental curve to a Gaussian-Lorentzian ratio variable curve using an iterative algorithm. Depth profiling analysis was obtained by sputtering the complete three-layer absorber with a 4.0 keV Argon ion flux at 1.5 eV acceleration voltage.
The condensation ageing experiment was performed in a Q.U.V. weathering chamber following the procedure developed by ISO 22975-3 (2014) standard [29]. The samples were subjected to condensation conditions at 40 • C for 150 h, 300 h and 600 h in order to determine the resistance to condensed water of the selective absorber. A thermal stability test was also carried out by using a conventional circulating air furnace where the samples were first heated at 250 • C for 200 h and then at 300 • C for 200 h, 400 h, 600 h and 800 h.
The optical properties of the absorbers were measured before and after the ageing studies. Regarding α s , this parameter was calculated as specified by the standard procedure [30] by using the direct AM1.5 solar spectrum from ASTM G173-03 in the 0.3-2.5 µm range [31] from solar hemispherical reflectance values (ρ s,h ). The hemispherical reflectance spectra of the samples were recorded from 0.3 to 2.5 µm with a UV-VIS-NIR Perkin-Elmer LAMBDA 950 double beam spectrophotometer equipped with a 150 mm Spectralon ® coated integrating sphere. The associated measurement uncertainty was 1%.
Samples ε T was also determined as specified by the standard procedure [29]. In this case, a Perkin Elmer Frontier FTIR spectrophotometer equipped with a diffuse goldcoated integrating sphere was used. A certified infragold standard was used as the reference material. The accuracy of the reflectance data was estimated to be 2% for the FTIR-spectrophotometer. Thermal emittance was calculated from the hemispherical IRreflectance spectra recorded from 2.5 to 17 µm measured at room temperature and the black body spectrum at 100 • C. For defining the acceptable durability or service life of the absorber, the performance criterion (PC) was determined. This parameter was calculated according to the ISO 22975-3 (2014) standard [29], that is, from the optical properties before and after the accelerated ageing test by applying Equation (1) in which the importance of the thermal emittance is reduced by a factor of 0.5: where ∆α s is the change in the solar absorptance defined as ∆α s = α s,t − α s,i being α s,t the solar absorptance at the actual time of the test or service, and α s,i the initial value of solar absorptance. ∆ε t is the change in the thermal emittance defined as ∆ε= ε t − ε i being ε t the thermal emittance at the actual time of the test or service, and ε i the initial value of thermal emittance.

Structural Study of the Al/CuMnO x /CuFeMnO x /SiO 2 Absorber
In order to study the crystalline structure and the composition of each constituent layer of the selective absorber, grazing angle XRD and XPS measurements were performed, respectively.

XRD Analysis
XRD patterns of the complete selective absorber were obtained where various peak characteristics of spinel phases could be identified. However, the interpretation of each layer's contribution to these spectra is very difficult due to peaks of spinel with different stoichiometry and composition occurring close to each other and the high intensity of the substrate (aluminium) peaks. Therefore, the XRD study of each film of the absorber was carried out separately. Figure 1 shows the XRD spectra obtained for Al/CuMnO x (a) and Al/CuFeMnO x (b) samples. reference material. The accuracy of the reflectance data was estimated to be 2% for the FTIR-spectrophotometer. Thermal emittance was calculated from the hemispherical IR-reflectance spectra recorded from 2.5 to 17 µm measured at room temperature and the black body spectrum at 100 °C.
For defining the acceptable durability or service life of the absorber, the performance criterion (PC) was determined. This parameter was calculated according to the ISO 22975-3 (2014) standard [29], that is, from the optical properties before and after the accelerated ageing test by applying Equation (1) in which the importance of the thermal emittance is reduced by a factor of 0.5: where Δαs is the change in the solar absorptance defined as Δαs= αs,t − αs,i being αs,t the solar absorptance at the actual time of the test or service, and αs,i the initial value of solar absorptance. Δεt is the change in the thermal emittance defined as Δε= εt − εi being εt the thermal emittance at the actual time of the test or service, and εi the initial value of thermal emittance.

Structural Study of the Al/CuMnOx/CuFeMnOx/SiO2 Absorber
In order to study the crystalline structure and the composition of each constituent layer of the selective absorber, grazing angle XRD and XPS measurements were performed, respectively.

XRD Analysis
XRD patterns of the complete selective absorber were obtained where various peak characteristics of spinel phases could be identified. However, the interpretation of each layer's contribution to these spectra is very difficult due to peaks of spinel with different stoichiometry and composition occurring close to each other and the high intensity of the substrate (aluminium) peaks. Therefore, the XRD study of each film of the absorber was carried out separately. Figure 1 shows the XRD spectra obtained for Al/CuMnOx (a) and Al/CuFeMnOx (b) samples.  These spectra show a clear crystallisation for both CuMnO x and CuFeMnO x films as indicated by the presence of characteristic peaks.
It should be noticed that the signals obtained for the films were less intense than the aluminium substrate ones. This is due to the larger proportion of aluminium substrate in the sample compared to the coatings whose thicknesses were only a few nanometers (70 nm for CuMnO x and 30 nm for CuFeMnO x ) [16]. Similar results were obtained by Amri et al. [32] who needed up to six dip-heating cycles in order to obtain spinel peaks of similar intensity as the aluminium peaks. For this reason and in order to ensure a clear visualisation of the XRD patterns related to the constituent absorber coatings, the diffractogram of the aluminium substrate is incorporated in each graphical representation. Concerning the peaks of the aluminium substrate, the identification was carried out by using the reference codes ICDD 00-001-1179 and ICDD 03-065-2869.
Based on the XRD pattern of the CuMnO x coating shown in Figure 1a, the first constituent absorber layer is composed of two coexisting crystalline phases: the spineltype structure of Cu 1 . 5 Mn 1 . 5 O 4 (ICDD 01-070-0262) and a secondary CuMn 2 O 4 spinel phase (ICDD 01-076-2296). Despite both crystalline phases coexisting in the CuMnO x film, the CuMn 2 O 4 phase is expected to be a minority in this coating compared to the Cu 1 . 5 Mn 1 . 5 O 4 spinel-type phase since only one weak diffraction peak at 43 • is assigned to the CuMn 2 O 4 spinel in the diffractogram. All other peaks, (discarding the peaks allocated to the aluminium) are assigned to the Cu 1 . 5 Mn 1 . 5 O 4 spinel phase.
In the same manner and in accordance with the ICDD database, the XRD results concerning the CuFeMnO x film indicate a mixture of two crystalline phases. As it is shown in Figure 1b, both coexisting phases correspond to the spinel-type structure of CuFeMnO 4 (ICDD 00-020-0358) and the magnetite phase Fe 3 O 4 (ICDD 00-001-1111).
In the case of SiO 2 coating, no crystalline structure was identified in the XRD analysis for any of the different thermal treatments studied, that is, blower for 5 min or oven for 15-30 min. These results were expected due to the soft thermal treatments to which the SiO 2 anti-reflective coating has been subjected for its sintering stage.
Definitely, XRD spectra confirmed the co-existence of two spinel-type phases in the first constituent absorber layer (Cu 1 . 5 Mn 1 . 5 O 4 and CuMn 2 O 4 ) and the co-existence of both CuFeMnO 4 spinel-type and Fe 3 O 4 magnetite phases in the second constituent absorber layer. Meanwhile, the SiO 2 over-coating was found to be amorphous. In no case peaks corresponding to alumina appeared, indicating that this layer formed during the sintering of each constituent layer was amorphous.

XPS Analysis
In order to confirm and complete the results obtained by XRD, an additional study by XPS was carried out on the CuMnO x , CuFeMnO x and SiO 2 coatings. This study allowed us to determine the element-binding states and the composition of the three constituent absorber layers.
• CuMnO x For the CuMnO x deposited layer, a metallic atomic ratio of Cu (16.6%)/Mn (16.8%) = 0.99 was obtained in the XPS analysis. These values are in accordance with the selected metallic ratio in the precursor solution (Cu:Mn 1:1). Additionally, the XPS technique also provided information about the oxidation state of Cu and Mn ions by studying their corresponding 2p 3/2 core level peak. Figure 2 shows the XPS spectra of Cu 2p (a), Mn 2p (b) and the congruent fit curves of Cu 2p 3/2 (c) and Mn 2p 3/2 (d) peaks recorded for the CuMnO x film annealed at 600 • C. As can be seen, the peaks associated with the metals present in the film are significantly complex. Therefore, contributions of various oxidation states are expected at first sight. Based on the Cu ion contribution in the CuMnO x film, four differentiated peaks in the deconvolution curve of Cu 2p 3/2 are observed. The peak at the low binding energy of 930.0 eV can be assigned to the Cu + ion in the Cu 1 . 5 Mn 1 . 5 O 4 spinel structure [33,34]. A second peak at 931.8 eV could be assigned to metallic Cu [35] taking into account its Cu LMM kinetic energy (918.0 eV, spectrum not shown here). Finally, two more peaks at 933.3 and 935.5 eV can be referred to the Cu 2+ spectrum [33,36]. The large shake-up satellite peak registered between 938 and 945 eV in the Cu2p spectrum ( Figure 2a) confirms the presence of Cu 2+ . The identification of peaks related to the same Cu 2+ species but at different positions is attributable to the existence of a different chemical environment of this ion in the spinel crystal lattice since these ions can occupy both octahedral and tetrahedral sites [10]. Specifically, these different contributions of the Cu 2+ ion in the CuMnO x film support the results obtained in the XRD analysis in which two types of spinel phases were detected: Cu 1 . 5 [11,34,37,38]. Additionally, by analysing the quantified results shown in Figure 2c, it can be seen that the Cu 2+ dominates in the CuMnO x film against the Cu + species. Regarding the second constituent species of the CuMnO x film, it is observed that the deconvolution curve of Mn 2p 3/2 is composed of three peaks at binding energies of 640.0, 641.5 and 642.9 eV corresponding to Mn 2+ , Mn 3+ and Mn 4+ [17,[38][39][40], respectively, and multiplet split components associated with Mn oxides at binding energies higher than 644 eV [17,38,39]. Regarding the site distribution of manganese ions, the majority of works reported in the literature agree that Mn 3+ and Mn 4+ ions occupy preferably octahedral sites and Mn 2+ is found in tetrahedral sites [41,42]. Based on the quantified results of the XPS analysis, which are summarised in Table Figure 3 shows the XPS spectra of Cu 2p (a), Mn 2p (b), Fe 2p (c) and the corresponding fit curves of Cu 2p 3/2 (d), Mn 2p 3/2 (e) and Fe 2p 3/2 (f) peaks recorded for the film annealed at 600 • C. Concerning the Cu element, by performing peak-fitting deconvolution, the Cu 2p 3/2 spectrum can be separated into two peaks: 932.7 and 934.8 eV, which correspond to the Cu 2+ situated in octahedral and tetrahedral sites, respectively [38]. Apart from the information given by the peaks at the definite binding energies, the satellite peak at 938-945 eV in the XPS spectrum of Cu 2p ( Figure 3a) also confirms the presence of Cu 2+ in the layer under study. Therefore, it can be confirmed that only one oxidation state of copper (Cu 2+ ) is detected in the CuFeMnO x layer, preferably, occupying octahedral sites (85%) than tetrahedral sites (15%) as the quantified results of the XPS analysis corroborate. In relation to the manganese ion contribution, a co-existence of various oxidation states is expected for this ion given the complexity of the registered peaks. In fact, in Figure 3e it can be seen that the deconvolution curve of Mn 2p 3/2 is composed of two peaks, namely, 640.8 and 642.7, corresponding to Mn 2+ and Mn 4+ , respectively, and multiplet-split components associated with Mn oxides at binding energies higher than 644 eV. As mentioned in previous paragraphs and in accordance with the majority of the references, the Mn 2+ is preferably found seizing tetrahedral sites whereas the Mn 4+ has a tendency to occupy octahedral sites. Based on these results, no differences in terms of the oxidation states of manganese are observed between both CuMnO x and CuFeMnO x films. However, by analysing and comparing the quantified results obtained for the manganese species present in each CuMnO tively, and multiplet split components associated with Mn oxides at binding en gies higher than 644 eV [17,38,39]. Regarding the site distribution of manganese io the majority of works reported in the literature agree that Mn 3+ and Mn 4+ ions o cupy preferably octahedral sites and Mn 2+ is found in tetrahedral sites [41,42]. Bas on the quantified results of the XPS analysis, which are summarised in   statement is corroborated by analysing the quantified results obtained in Figure  that is, 50% for Fe 3+ ions occupying octahedral sites while 24% for Fe 3+ ions occup ing tetrahedral sites. Certainly, both XRD and XPS results support the formation the spinel-like material of the FeMnCuO4 phase, as well as the presence of the se regated oxide of iron Fe3O4 in the second constituent absorber layer. The data related to the oxidation-state percentage and site distribution of copp manganese and iron ions in CuFeMnOx thin film are summarised in Table 2. The data related to the oxidation-state percentage and site distribution of copper, manganese and iron ions in CuFeMnO x thin film are summarised in Table 2. • SiO 2 Since the XRD analysis of the SiO 2 layer reports the formation of an amorphous phase of the silica film at the studied sintering conditions, the composition of this third layer was studied by using the XPS technique. Figure 4 shows the XPS survey spectrum for the silica layer sintered in the blower and the corresponding Si 2p spectrum (inset). Based on this graphical representation, it can be seen that the protective and antireflective SiO 2 film is composed of Si and O. The presence of C in the spectrum is associated with the environmental contamination of the sample. According to these results, it can be noticed that the entire silicon ion in the layer is present as Si 4+ , being, therefore, part of the SiO 2 layer in its entirety. • SiO2 Since the XRD analysis of the SiO2 layer reports the formation of a phase of the silica film at the studied sintering conditions, the compo third layer was studied by using the XPS technique. Figure 4 shows th spectrum for the silica layer sintered in the blower and the corresp spectrum (inset). Based on this graphical representation, it can be seen tective and antireflective SiO2 film is composed of Si and O. The presen spectrum is associated with the environmental contamination of the cording to these results, it can be noticed that the entire silicon ion i present as Si 4+ , being, therefore, part of the SiO2 layer in its entirety.

Effect of the SiO2 Sintering Process on the Optical Behaviour
In this section, a detailed study of the influence of the SiO2 sintering p optical parameters of the proposed absorber Al/CuMnOx/CuFeMnOx/SiO2 The SiO2 sintering was performed by using either a blower for 5 min or an 30 min. In Figure 5 it can be seen how the variation of this SiO2 sintering ences the hemispherical reflectance spectra of the complete absorber samp higher values of reflectance in the second maximum (from 0.70 µm to th edge) are reached for the absorber samples whose third layer was sintere during both 15 and 30 min (7-8% at 1.15 µm) instead of using the blower (3% In addition, the absorbers whose SiO2 layer was sintered in the oven had t edge shifted to higher wavelengths and showed a decrease in reflectanc range. Specifically, the absorption edge appears at around 1.60 µm for whose silica layer was sintered in the oven while for the sample whose S sintered in the blower the absorption edge is placed at 1.45 µm. Moreov

Effect of the SiO 2 Sintering Process on the Optical Behaviour
In this section, a detailed study of the influence of the SiO 2 sintering process on the optical parameters of the proposed absorber Al/CuMnO x /CuFeMnO x /SiO 2 is presented. The SiO 2 sintering was performed by using either a blower for 5 min or an oven for 15 or 30 min. In Figure 5 it can be seen how the variation of this SiO 2 sintering process influences the hemispherical reflectance spectra of the complete absorber samples. Basically, higher values of reflectance in the second maximum (from 0.70 µm to the absorption edge) are reached for the absorber samples whose third layer was sintered in the oven during both 15 and 30 min (7-8% at 1.15 µm) instead of using the blower (3% at 1.15 µm). In addition, the absorbers whose SiO 2 layer was sintered in the oven had the absorption edge shifted to higher wavelengths and showed a decrease in reflectance in the NIR range. Specifically, the absorption edge appears at around 1.60 µm for both samples whose silica layer was sintered in the oven while for the sample whose SiO 2 layer was sintered in the blower the absorption edge is placed at 1.45 µm. Moreover, it was observed that as the sintering time in the oven increased, the second maximum reached slightly higher values of reflectance accompanied by a decrease in the reflectance in the NIR range. These variations in the curve shape of the Al/CuMnOx/CuFeM strongly suggest that thicker layers are obtained in case of sintering t oven instead of sintering in the blower. This thicker layer resulting process in the oven could be attributed to oxide layer formation betw substrate and the absorber as a consequence of the oxidation of the al at longer sintering times. This fact is confirmed in Figure 6, which sh spectra of the bare aluminium substrate and of the SiO2 coated al (Al/SiO2) both treated under the same conditions. In the bare alumini decreases with the increase in the sintering time (300-800 nm) sugg layer formation. When the SiO2 is deposited, the same effect can be ob  These variations in the curve shape of the Al/CuMnO x /CuFeMnO x /SiO 2 samples, strongly suggest that thicker layers are obtained in case of sintering the SiO 2 layer in the oven instead of sintering in the blower. This thicker layer resulting from the sintering process in the oven could be attributed to oxide layer formation between the aluminium substrate and the absorber as a consequence of the oxidation of the aluminium substrate at longer sintering times. This fact is confirmed in Figure 6, which shows the reflectance spectra of the bare aluminium substrate and of the SiO 2 coated aluminium substrate (Al/SiO 2 ) both treated under the same conditions. In the bare aluminium, the reflectance decreases with the increase in the sintering time (300-800 nm) suggesting the alumina layer formation. When the SiO 2 is deposited, the same effect can be observed. These variations in the curve shape of the Al/CuMnOx/C strongly suggest that thicker layers are obtained in case of sinter oven instead of sintering in the blower. This thicker layer resu process in the oven could be attributed to oxide layer formation substrate and the absorber as a consequence of the oxidation of t at longer sintering times. This fact is confirmed in Figure 6, whic spectra of the bare aluminium substrate and of the SiO2 coate (Al/SiO2) both treated under the same conditions. In the bare alu decreases with the increase in the sintering time (300-800 nm) layer formation. When the SiO2 is deposited, the same effect can b Figure 6. Effect of the SiO2 sintering conditions on the bare Al substrate a hemispherical reflectance measurements. All the curve shape variations described in both Figure 5 a expected, changes in the absorptance values of the samples. As significant loss of the solar absorptance value can be found by sin the oven (0.950-0.953) compared to using the blower (0.957). All the curve shape variations described in both Figures 5 and 6 implied, as expected, changes in the absorptance values of the samples. As is shown in Table 3, a significant loss of the solar absorptance value can be found by sintering the silica layer in the oven (0.950-0.953) compared to using the blower (0.957).

Effect of the SiO 2 Sintering Process on Durability
Apart from good optical properties, which have already been discussed in previous work for the proposed three-layer configuration absorber Al/CuMnO x /CuFeMnO x /SiO 2 , absorber coatings have to accomplish long-term stability during service life as the microclimatic conditions can significantly affect the performance of the material. The effect of the SiO 2 sintering conditions on durability has been also studied since this layer is often used as a protective layer as well as an anti-reflective [46]. Thermal stability tests and condensation tests were carried out separately for each sample.
Regarding the thermal stability test, six replicas of each three-layer configuration sample whose SiO 2 layer was sintered in the blower (5 min) or in the oven (15 min), were heated at 250 • C for 200 h and afterwards at 300 • C for 200 h, 400 h, 600 h and 800 h. After each degradation time interval, the samples were first visually inspected and no apparent deteriorations were observed. All the samples kept their initial appearances and the colour remained unchanged. As an example, photographs of the Al/CuMnO x /CuFeMnO x /SiO 2 (blower) selective absorber before (a) and after (b) the thermal stability study are shown in Figure 7. Figure 8 displays the hemispherical reflectance spectra for the sample Al/CuMnO x /CuFeMnO x /SiO 2 whose silica layer was sintered in the blower, before and after each ageing stage for the wavelength range 0.3-2.5 µm (Figure 8a) and 0.3-17 µm (Figure 8b). Both graphical representations show hardly any difference between the spectra, corroborating an almost non-existent modification of the silica films during the thermal stability test. It should be emphasised that this effect was also observed for the rest of the samples whose SiO 2 layer was sintered in the oven for 15 min.

Effect of the SiO2 Sintering Process on Durability
Apart from good optical properties, which have already work for the proposed three-layer configuration absorber Al absorber coatings have to accomplish long-term stability du cro-climatic conditions can significantly affect the performanc of the SiO2 sintering conditions on durability has been also st ten used as a protective layer as well as an anti-reflective [46]. condensation tests were carried out separately for each sample Regarding the thermal stability test, six replicas of each sample whose SiO2 layer was sintered in the blower (5 min) or heated at 250 °C for 200 h and afterwards at 300 °C for 200 h, 4 each degradation time interval, the samples were first visually deteriorations were observed. All the samples kept their in colour remained unchanged.
As an example, Al/CuMnOx/CuFeMnOx/SiO2 (blower) selective absorber be thermal stability study are shown in Figure 7. Figure 8 disp flectance spectra for the sample Al/CuMnOx/CuFeMnOx/SiO2 tered in the blower, before and after each ageing stage for the μm (Figure 8a) and 0.3-17 μm (Figure 8b). Both graphical re any difference between the spectra, corroborating an almost n the silica films during the thermal stability test. It should be was also observed for the rest of the samples whose SiO2 laye for 15 min.    Seeing minimal variations in the hemispherical reflectance spectra, neither notable fluctuations in both solar absorptance and thermal emittance values were expected for the tested samples. Indeed, Table 4 summarises the negligible variations observed in the optical parameters of the samples before and after each exposure period as well as the resulting PC value. Since the influence of ageing on the efficiency of the collector must be less than 5% after 25 years, the samples have to reach PC values lower than 0.05 at the end of the test to be accepted, as the standard specifies [29]. Values in Table 4 show that the optical performance of all the absorbers studied is reduced less than 5% of its original value during the design service life time period. In fact, each sample demonstrated to be very stable, with the PC values registered in all cases close to zero. As a result of these very small PC values, it can be considered that all the studied absorber surfaces are qualified as far as thermal stability is concerned, even after approximately 1000 h of testing, regardless of the sintering process of the SiO 2 layer. Therefore, the excellent thermal stability results obtained indicate that the constituent absorber coatings are suitable for solar absorbers in terms of thermal stability performance.
Following the long term durability study of the absorbers with different SiO 2 sintering processes, a condensation test was performed. As a general trend, condensation tests are considered much more aggressive than thermal stability tests. This condensation test involved the exposure of six replicas of each tested absorber under constant condensation (40 • C) in the weathering chamber with interruptions to measure the extent of degradation after 150 h, 300 h and 600 h. At each interruption, the samples were first examined by visual inspection for evaluating the absorber coatings and then the optical measurements were performed to calculate the solar absorptance, thermal emittance and PC values.
Due to the combined effect of condensation, high humidity and temperature (40 • C), some colour changes and sample degradation were observed for the three types of aluminium based solar selective absorbers studied, as shown in Figure 9. Particularly, pitting corrosion and a minor colour degradation after the first 150-300 h of testing were observed for the samples analysed. After each testing time in the weathering chamber, the reflectance s samples were also measured and the absorbers' optical performance was Figure 10 shows the variations registered in the hemispherical reflectan Al/CuMnOx/CuFeMnOx/SiO2 (blower) and Al/CuMnOx/CuFeMnOx/SiO2 (o samples before and after each ageing step in the weathering chamber. Bot tions are selected in this manuscript due to the interesting conclusions re study, being clearly visible the variations produced in the spectra during variations observed occur mainly in the solar wavelength range of 0.3-2 specific case of the Al/CuMnOx/CuFeMnOx/SiO2 (blower) sample (Figure 1 served that the first maximum (around 0.4 µm) reaches lower values of refle 6.8%) as the condensation test progresses. In addition, an increase of the mum (0.80 µm until the absorption edge), a displacement of the absorp higher wavelengths and a decrease in the NIR range are also observed. The as the test time proceeds are characteristic of an increase in the layer th However, as this effect is improbable in these conditions, the variations ob curve shape of the Al/CuMnOx/CuFeMnOx/SiO2 (blower) samples after the test could be attributed to the oxidation of the aluminium substrate. On the After each testing time in the weathering chamber, the reflectance spectra of the samples were also measured and the absorbers' optical performance was determined. Figure 10 shows the variations registered in the hemispherical reflectance spectra of Al/CuMnO x /CuFeMnO x /SiO 2 (blower) and Al/CuMnO x /CuFeMnO x /SiO 2 (oven 30 min) samples before and after each ageing step in the weathering chamber. Both representations are selected in this manuscript due to the interesting conclusions reached in this study, being clearly visible the variations produced in the spectra during the test. The variations observed occur mainly in the solar wavelength range of 0.3-2.5 µm. In the specific case of the Al/CuMnO x /CuFeMnO x /SiO 2 (blower) sample (Figure 10a)), it is observed that the first maximum (around 0.4 µm) reaches lower values of reflectance (10.0-6.8%) as the condensation test progresses. In addition, an increase of the second maximum (0.80 µm until the absorption edge), a displacement of the absorption edge to higher wavelengths and a decrease in the NIR range are also observed. These variations as the test time proceeds are characteristic of an increase in the layer thickness [16]. However, as this effect is improbable in these conditions, the variations observed in the curve shape of the Al/CuMnO x /CuFeMnO x /SiO 2 (blower) samples after the condensation test could be attributed to the oxidation of the aluminium substrate. On the other hand, a different tendency is observed in the Al/CuMnO x /CuFeMnO x /SiO 2 (oven 30 min) samples (Figure 10b)). In this case, the first maximum (around 0.4 µm) of the sample spectrum reaches higher values of reflectance as the condensation test progresses (7.5-11%).
Meanwhile the second maximum (0.80 µm until the absorption edge) maintains nearly its original position around the reflectance value of 8.5%. Apart from these variations, it is also observed a slight shift of the absorption edge to lower wavelength values as the test progresses. Definitely, these variations in the shape of the Al/CuMnO x /CuFeMnO x /SiO 2 (oven 30 min) samples spectra suggest that the silica antireflective layer becomes thinner as the test in the weathering chamber progresses, that is, the SiO 2 over-coating degrades along the condensation test. Optical parameters as well as the resulting PC values registered for each sample before and after each cycle of the condensation test are recorded in Table 5. According to optical parameters analysis, a slight decrease of the solar absorptance value and an increase of the thermal emittance value with the exposure were observed for all the Al/CuMnO x /CuFeMnO x /SiO 2 samples studied. There is an exception for exposure times between 150 and 300 h in which the values of solar absorptance, thermal emittance and PC remained unaltered. In spite of the changes observed in the optical properties values and aspects of the samples after the condensation test, it can be concluded that all the absorbers studied registered a final PC value well below the critical value for which an absorber can be considered qualified, that is, the absorbers' optical performance was still good for all samples with a PC ≤ 0.015 (Table 5) after 600 h of testing at 40 • C.
Additional comparative experiments were carried out in order to delve into the role that the alumina layer produced during the preparation of the three-layer complete absorber and the silica top layer plays in protecting the samples against constant condensation conditions. For this purpose, preheated Al at 600 • C (simulating the sintering treatments of the complete absorber) and not-preheated were used to apply the upper silica layer and were tested in the condensation chamber (three replicas). Figure 11 shows the hemispherical reflectance spectra of not-preheated Al/SiO 2 samples before and after 147 h of condensation test. The reflectance decreases considerably (around 8%) for all the samples and these variations demonstrate the degradation that the three samples undergo at the condensation conditions tested. The inset in Figure 11 also shows the variation of the hemispherical reflectance of bare aluminium before and after 147 h of condensation test. It should be noted that there is a drastic reflectance decrease, from 89% to 37% after this testing time. Figure 12 shows the photos of the samples before and after 147 h of testing. It can be observed that the SiO 2 layer protects to a great extent the Al substrate from oxidation but not enough to avoid being unaltered. The Al/SiO 2 samples exhibited clear degradation, going from being transparent and specular before the condensation test to being whitish and matte after 147 h of testing. On the other hand, the bare Al appeared completely oxidised and degraded in the same conditions.
Additional comparative experiments were carried out in order that the alumina layer produced during the preparation of the thre sorber and the silica top layer plays in protecting the samples again sation conditions. For this purpose, preheated Al at 600 °C (simu treatments of the complete absorber) and not-preheated were used silica layer and were tested in the condensation chamber (three replic the hemispherical reflectance spectra of not-preheated Al/SiO2 samp 147 h of condensation test. The reflectance decreases considerably (a samples and these variations demonstrate the degradation that th dergo at the condensation conditions tested. The inset in Figure 11 a tion of the hemispherical reflectance of bare aluminium before and densation test. It should be noted that there is a drastic reflectance d 37% after this testing time. Figure 12 shows the photos of the samples h of testing. It can be observed that the SiO2 layer protects to a gre strate from oxidation but not enough to avoid being unaltered. The hibited clear degradation, going from being transparent and spec densation test to being whitish and matte after 147 h of testing. On bare Al appeared completely oxidised and degraded in the same con    The hemispherical reflectance spectra of Al/SiO2 preheated before a layer together with bare aluminium preheated and heated (simulat treatment of the SiO2 layer) are plotted in Figure 13. The preheating The hemispherical reflectance spectra of Al/SiO 2 preheated before applying the silica layer together with bare aluminium preheated and heated (simulating the sintering treatment of the SiO 2 layer) are plotted in Figure 13. The preheating treatment of the aluminium substrate induces greater variations in the reflectance spectra between the different SiO 2 sintering conditions. Furthermore, better results regarding durability in the climate chamber were obtained in comparison with the non-pre-heated Al samples. In fact, the pre-heating of Al/SiO 2 samples at 600 • C allowed all the analysed samples to remain intact after 600 h of condensation testing regardless of the SiO 2 sintering process. The solar reflectance data after each degradation interval during the condensation test of the preheated Al/SiO 2 samples are reported in Table 6. All the replicas tested duplicate the pristine hemispherical reflectance spectra. These results confirm that the good resistance to the humidity of the proposed absorber is due to the protective action of both the silica layer and alumina layer produced naturally during the sintering process of the whole absorber preparation. The antireflective SiO 2 layer protects, to some extent, the aluminium from oxidation but is not enough to withstand the condensation test conditions. Materials 2022, 15, x FOR PEER REVIEW aluminium substrate induces greater variations in the reflectance spectra betw different SiO2 sintering conditions. Furthermore, better results regarding durabilit climate chamber were obtained in comparison with the non-pre-heated Al sam fact, the pre-heating of Al/SiO2 samples at 600 °C allowed all the analysed sam remain intact after 600 h of condensation testing regardless of the SiO2 sintering p The solar reflectance data after each degradation interval during the condensation the preheated Al/SiO2 samples are reported in Table 6. All the replicas tested du the pristine hemispherical reflectance spectra. These results confirm that the g sistance to the humidity of the proposed absorber is due to the protective action the silica layer and alumina layer produced naturally during the sintering proces whole absorber preparation. The antireflective SiO2 layer protects, to some ext aluminium from oxidation but is not enough to withstand the condensation tes tions. Figure 13. Effect of SiO2 sintering conditions on both preheated bare Al and Al/SiO2 on th spherical reflectance spectra.

Sintering
Solar Reflectance (ρs,h) Figure 13. Effect of SiO 2 sintering conditions on both preheated bare Al and Al/SiO 2 on the hemispherical reflectance spectra. Even though the three kinds of absorbers studied can be qualified with respect to their resistance to high temperatures (stability test) and constant condensation conditions (condensation test), it would be interesting to analyse how the sintering conditions of the silica can affect the wetting ability of the absorber surface. For this reason, static contact angle measurements were carried out before subjecting the samples to any test and the results are recorded in Table 7. It can be observed from the measurements that the different silica sintering processes studied affect differently the wetting ability of the absorber surface. In particular, the static contact angle value decreases from 37 • to 25-21 • depending on whether the SiO 2 anti-reflective coating has been sintered in the blower or oven. Specifically, the highest static contact angle value was obtained for the absorber whose silica layer was sintered in the blower while the lowest static contact angle value was registered for the absorber whose SiO 2 layer was sintered in the oven for 30 min. These results support that the longer the SiO 2 layer is sintered in the oven, the more organic matter is burnt making the layer more hydrophilic and then the moisture of the condensation test stays longer in the sample as the surface becomes wetter. In conclusion, based on the results obtained in both degradation tests of thermal stability and condensation resistance (absorptance values, thermal emittance and PC ≤ 0.015), it can be confirmed that the CuMnO x /CuFeMnO x films were effectively protected by the combined action of the alumina layer produced onto the aluminium substrate during the absorber preparation and the SiO 2 over-coating; regardless of the sintering process used for the SiO 2 layer. In this way, it can be summarised that optimal results were obtained for any of the samples studied.
Once both durability and characterisation studies have been completed, a depth profile study of the selective absorber Al-CuMnO x /CuFeMnO x /SiO 2 whose silica layer was sintered in the blower was carried out by using XPS. This three layer-configuration sample was successfully etched with an Ar + ion beam. Figure 14 shows the results of the depth profile study of this configuration before (a) and after (b) the ageing tests. It can be seen that the silicon concentration decreases gradually as the etching proceeds whereas the copper, manganese and iron concentrations firstly increase, attain saturation and finally decrease gradually with etching time. Immediately after the etching reaches the Al substrate, an increase in the concentration of the Al, together with an increase in the oxygen content, are observed in both pristine and aged samples. This fact confirms the formation of the alumina layer onto the aluminium substrate during the absorber sintering. Additionally, in aged samples, (Figure 14b), a hydration phenomenon of spinel components was perceived as a consequence of the constant humidity conditions to which the samples have been subjected. The iron spinel stops appearing 28 min after sputtering in the case of the pristine samples while in the case of the tested samples the iron stops appearing after approximately 40 min of sputtering.

Conclusions
In this paper, the structural characterisation of the constituent absorber films was studied. The XRD and XPS results, confirm the formation of spinel-like materials in the first absorber layer with two different stequiometries, Cu1.5Mn1.5O4 being the main phase. With regards to the second layer, the co-existence of the spinel CuFeMnO4 with a secondary magnetite phase of Fe3O4 has been demonstrated. Additionally, durability tests of this competitive three-layer configuration solar selective absorber prepared by dip-coating have been presented. The results obtained showed that the sintering conditions of the SiO2 top layer affect the optical properties as well as the accelerated ageing test results. In general, all the sintering conditions used gave place to selective absorbers with very good optical performance (αs ≈ 0.957 and εt (100 °C) ≈ 0.038) and excellent durability results. Specifically, PC values close to 0.001 in the thermal stability test and below 0.015 in the condensation test were obtained. On the other side, it was shown that the protective action against condensation results from the combination of both alumina layers produced during the absorber layer preparation and the antireflective silica layer. A hydration phenomenon of Fe spinel is confirmed by XPS for tested samples as a consequence of the constant humidity conditions to which the samples are subjected during the climatic chamber test. The results obtained in this work validate the proposed structure Al/CuMnOx/CuFeMnOx/SiO2 as a very promising selective absorber for low temperature solar thermal applications. The XPS depth profile results demonstrate the formation of the protective alumina layer during the sintering of the spinel layers and, on the other hand, they clearly suggest that the variations produced in the optical properties of samples tested under condensation could be produced by a hydration process of the CuFeMnO x intermediate films.

Conclusions
In this paper, the structural characterisation of the constituent absorber films was studied. The XRD and XPS results, confirm the formation of spinel-like materials in the first absorber layer with two different stequiometries, Cu 1 . 5 Mn 1 . 5 O 4 being the main phase. With regards to the second layer, the co-existence of the spinel CuFeMnO 4 with a secondary magnetite phase of Fe 3 O 4 has been demonstrated. Additionally, durability tests of this competitive three-layer configuration solar selective absorber prepared by dip-coating have been presented. The results obtained showed that the sintering conditions of the SiO 2 top layer affect the optical properties as well as the accelerated ageing test results. In general, all the sintering conditions used gave place to selective absorbers with very good optical performance (α s ≈ 0.957 and ε t (100 • C) ≈ 0.038) and excellent durability results. Specifically, PC values close to 0.001 in the thermal stability test and below 0.015 in the condensation test were obtained. On the other side, it was shown that the protective action against condensation results from the combination of both alumina layers produced during the absorber layer preparation and the antireflective silica layer. A hydration phenomenon of Fe spinel is confirmed by XPS for tested samples as a consequence of the constant humidity conditions to which the samples are subjected during the climatic chamber test. The results obtained in this work validate the proposed structure Al/CuMnO x /CuFeMnO x /SiO 2 as a very promising selective absorber for low temperature solar thermal applications.